J Gastrointest Surg (2009) 13:1358–1367

DOI 10.1007/s11605-009-0882-y

ORIGINAL ARTICLE

Identification of Patients at Risk for Development of Tertiary

Peritonitis on a Surgical Intensive Care Unit
Ansgar M. Chromik & Andreas Meiser & Janine Hölling & Dominique Sülberg &
Adrien Daigeler & Kirsten Meurer & Heike Vogelsang & Matthias H. Seelig &
Waldemar Uhl

Received: 31 December 2008 / Accepted: 24 March 2009 / Published online: 8 April 2009
# 2009 The Society for Surgery of the Alimentary Tract

Abstract
Background Tertiary peritonitis (TP) is defined as a severe recurrent or persistent intra-abdominal infection after adequate
surgical source control of secondary peritonitis (SP). The aim of this study was to analyze the characteristics of patients with
SP who will further develop TP in order to define early diagnostic markers for TP.
Study Design Over a 1-year period, all patients on the surgical intensive care unit (ICU) with SP were prospectively
assessed for the development of TP applying the definition of the ICU consensus conference. The Mannheim Peritonitis
Index (MPI), C-reactive protein (CRP) and Simplified Acute Physiology Score II (SAPS II) were assessed at the initial
operation (IO) that was diagnostic for SP and in the postoperative period.
Results Among 69 patients with SP, 15 patients further developed TP, whereas 54 patients did not develop TP. Compared to
SP, patients with transition to TP had significantly higher MPI at IO (28.6 vs. 19.8; p<0.001), relaparotomy rate (2.00 vs.
0.11; p<0.001), mortality (60% vs. 9%; p<0.001), duration of ICU stay (14 vs. 4 days; p<0.005), as well as SAPS II (45.1
vs. 28.4; p<0.005) and CRP (265 mg/dL vs. 217 mg/dL; p<0.05) on the second postoperative day after IO.
Conclusions The MPI at IO as well as CRP and SAPS II at the second postoperative day helps to identify patients at risk for
tertiary peritonitis.

Keywords C-reactive protein . SAPS II . Abbreviations

Mannheim Peritonitis Index . Sepsis . Secondary peritonitis CRP C-reactive protein
ICU Intensive care unit
This paper was presented by the principal author (A.M. Chromik) at IO Initial operation
the “Papers Session GS71 Trauma and Critical Care” of the 94th MPI Mannheim Peritonitis Index
Annual Clinical Congress of the American College of Surgeons,
SAPS II Simplified Acute Physiology Score II
October 12–16, 2008, San Francisco, (Abstract # NP2008-1246).
SP Secondary peritonitis
Ansgar M. Chromik and Andreas Meiser contributed equally to this TP Tertiary peritonitis
publication.
SP patient Patient with SP who did not further
A. M. Chromik : J. Hölling : D. Sülberg : K. Meurer : develop TP
M. H. Seelig : W. Uhl (*)
TP patient Patient with SP who further developed TP
Department of General and Visceral Surgery, St. Josef-Hospital,
Ruhr-University Bochum, Gudrunstrasse 56,
44791 Bochum, Germany
e-mail: w.uhl@klinikum-bochum.de Introduction
A. Meiser : H. Vogelsang
Department of Anesthesiology, St. Josef-Hospital, Definition of Tertiary Peritonitis
Ruhr-University Bochum, Bochum, Germany
Peritonitis is one of the most frequent diagnoses on a
A. Daigeler
Department of Plastic Surgery, Burn Center, Hand Center,
surgical intensive care unit leading to severe sepsis.1 It is
Sarcoma Reference Center, BG-University Hospital defined as an intra-abdominal peritoneal infection and can
Bergmannsheil, Ruhr-University Bochum, Bochum, Germany be classified into three major groups—primary, secondary,
J Gastrointest Surg (2009) 13:1358–1367
and tertiary peritonitis. Primary peritonitis—also referred to
as spontaneous bacterial peritonitis—arises in the absence
of an identifiable anatomical derangement and has a low
incidence on surgical intensive care units. The most
frequent entity is secondary peritonitis (SP) which is
defined as an infection of the peritoneal cavity resulting
from perforation, anastomotic disruption, ischemic necrosis,
or other injuries of the gastrointestinal tract.2 Operative
therapy is the treatment of choice and comprises surgical
source control of the infectious focus and reduction of the
bacterial load. Tertiary peritonitis (TP) is less common and
is defined as a severe recurrent or persistent intra-
abdominal infection after apparently successful and ade-
quate surgical source control of SP.2 It is characterized by a
prolonged systemic inflammation and organ dysfunction
leading to a high rate of SIRS, sepsis, severe sepsis, or
septic shock.1,3 As a result, mortality of TP ranges between
30% and 64%.2,4,5 The microbial flora encountered in TP is
different from SP and displays mostly opportunistic and
nosocomial facultative pathogenic bacteria and fungi (e.g.,
Enterococci, Enterobacter, Candida). Due to broad-
spectrum antibiotic therapy, a significant proportion of
microbes develop multi-resistance to antibiotics.
Diagnosis of TP
It is often difficult to differentiate between SP and TP since
there is a continuum between both clinical situations and the
exact time point when SP turns into TP is often missed.
Figure 1 illustrates different clinical scenarios for patients with
SP. If SP is diagnosed during an operation—which is referred
to as “the initial operation” (IO) in this context—the patient
will receive surgical source control (e.g., Hartmann’s proce-
dure for colonic perforation). If surgical source control is
successful, the majority of patients will recover. However, a
subset of patients will develop clinical signs of recurrent or
persistent intra-abdominal infection in spite of apparently
successful source control, which often results in a reoperation.
During subsequent relaparotomies, recurrent or persistent
peritonitis is encountered in spite of adequate and successful
surgical source control during the IO. This form of peritonitis
is referred to as TP. Importantly, the diagnosis of TP can only
be made in the absence of an obvious anatomical defect or
disruption of the gastrointestinal hollow viscera; otherwise,
the peritonitis has to be classified as ongoing SP—character-
ized by a primary failure of surgical source control (e.g.,
breakdown of the closure of the Hartmann’s pouch or
breakdown of the suture repair following gastric perforation;
Fig. 1). In fact, the most frequent way to diagnose TP, is a
“planned” or “on demand” relaparotomy, which is performed
in the interval after the IO (Fig. 1).6,7 However, a relapar-
otomy—either “planned” or “on demand”—may represent a
late event in the management of peritonitis, and it is not
1359
Figure 1 The diagnostic criteria for tertiary peritonitis (TP) and the
diagnostic challenge.
necessarily the first relaparotomy after the IO when TP is
encountered. Therefore, timely—non-operative—diagnosis of
TP after the IO and subsequent initiation of an appropriate
therapy may help to reduce the complication rate and to
improve the prognosis. It is desirable to identify patients at
risk for developing TP as early as possible or at least during
the first days after the IO for SP.
Diagnostic Challenge
The value of clinical and laboratory parameters and scoring
systems for sufficient diagnosis and monitoring of TP is
still discussed controversially.5 However, the intensive care
unit (ICU) consensus conference provided three categories
for the diagnostic certainty of TP: “microbiologically
confirmed”, “probable”, and “possible”.2 The Mannheim
Peritonitis Index (MPI) represents a scoring system that
estimates the severity and prognosis of secondary peritoni-
tis at the onset of SP. It is applied easily under routine
conditions during initial surgery for SP in the operating
room. It was developed and first described in 1987 by
Linder et al.8 and validated in several studies for SP.9,10
Recent studies reported encouraging results for the Man-
nheim Peritonitis Index regarding detection patients at risk
for TP.11,12 Another score that has shown a potential to be
successfully applied in TP is the Simplified Acute Physi-
ology Score II (SAPS II) score.12 It was initially designed
to predict mortality and disease severity of critically ill
patients on surgical intensive care units.13,14 Laboratory
parameters like C-reactive protein or procalcitonin have
rarely been evaluated in the diagnosis of TP.5,15
However, there is still a lack of studies addressing the
identification of risk factors for patients prone to develop
TP. It would be desirable to have diagnostic markers that
could predict at the onset of peritonitis—during the initial
operation or the first postoperative days after—whether the
individual patient will develop TP or not (Fig. 1).
The aims of this study were therefore (1) to compare
patients’ characteristics, clinical outcome and microbial flora
of patients with SP and TP and (2) to investigate the efficacy
of clinical and laboratory parameters like C-reactive protein,
1360
Mannheim Peritonitis Index and SAPS II to early identify
patients with SP at risk for the development of TP.
Material and Methods
Study Population and Definition of Secondary/Tertiary
Peritonitis
During a 1-year period (01.01.2006–31.12.2006), all patients
admitted to the surgical intensive care unit with a SP were
recorded in a prospective database. Due to hospital policy, all
patients with a secondary peritonitis are mandatorily admitted
to the surgical intensive care unit—for at least 24 h. SP had to
be diagnosed during a laparotomy, which was referred to as
the IO (Fig. 1). During follow-up, patients with SP were
continuously analyzed for the diagnosis of TP—in accor-
dance with the “International Sepsis Forum Consensus
Conference”.2 TP was therefore defined as intra-abdominal
infection that persists or recurs ≥48 h following successful
and adequate surgical source control during the IO.2 As
indicated in Fig. 1, patients with an obvious failure of
surgical source control after the IO or following procedures
(e.g., insufficiency of the rectal stump, anastomotic insuffi-
ciency, etc.) were considered as ongoing SP and not as TP.
Demographic data, origin of peritonitis and intra-
operative findings during IO, type of surgical procedure
performed during IO, antibiotic treatment, and follow-up
procedures like relaparotomies were collected. In order to
assess the severity of peritonitis as early as at the IO, the
Mannheim Peritonitis Index was calculated routinely during
the IO as previously described.8,9 Furthermore, C-reactive
protein was monitored daily during the first three postop-
erative days and on postoperative day 7 after IO. SAPS II
scores were recorded during the first three postoperative
days after IO as previously described.14 Mortality was
defined as any death during postoperative hospitalization.
Furthermore, intra-operative specimens of abdominal fluid
were analyzed by standard microbiological techniques.
Statistical Analysis
Results for the Mannheim Peritonitis Index were expressed as
median and displayed in box plots. Box plots are representing
the lower, median, and upper quartile whereas whiskers
indicate the 10th–90th percentile. Outliers are illustrated by
dots. Age, body mass index, Mannheim Peritonitis Index,
intensive care unit stay, and the number of relaparotomies
per patient were compared by Mann–Whitney test.
Frequencies for co-morbidities, underlying malignancy
as well as mortality data, frequency of relaparotomies and
frequency of specific bacteria were compared by Fisher’s
exact test. C-reactive protein values and SAPS II scores
J Gastrointest Surg (2009) 13:1358–1367
are expressed as means ± SEM and compared by T test. p
values ≤0.05 were considered statistically significant. To
determine the diagnostic accuracy of the Mannheim
Peritonitis Index measurement during initial operation as
well as C-reactive protein and SAPS II measurements
2 days after initial operation, for the distinction between
TP and SP, corresponding receiver operating characteristic
curves were calculated. Furthermore, the area under the
receiver operator characteristic curve was defined. Cut-off
values for the Mannheim Peritonitis Index, C-reactive
protein, and SAPS II with the corresponding sensitivity,
specificity, and confidence intervals were given. Data
were processed with SPSS 16.0/GraphPadPrism 5.
Results
Demographic Data of the Study Population
Over a 1-year period (2006), 1,091 patients were admitted to
the surgical intensive care unit. Among the 1,091 intensive
care unit patients, 69 were diagnosed having SP. The
diagnosis of SP was made intra-operatively in all 69 patients
during the IO. Among those, 15 patients (21.7%) further
developed TP—according to the ICU consensus conference
definition.2 These patients were referred to as TP patients
throughout this study. The remaining 54 patients with SP
(78.3%) did not develop TP and were therefore referred to as
SP patients. Demographic data of the study population are
summarized in Table 1. There was no significant difference
in gender distribution, age, body mass index, cardiovascular
and pulmonary co-morbidities as well as malignant diseases
between SP and TP patients (Table 1).
Etiology and Source of Peritonitis
Etiologies and infection source of secondary peritonitis for
all patients (n=69) found at the IO are depicted in Fig. 2,
separately for TP patients (n=15) and SP patients (n=54).
The majority of patients had perforated diverticulitis or
other colonic perforations at the IO. Less frequent were
other causes like gastric/duodenal perforations, anastomotic
insufficiencies, or appendicitis. However, there was no
significant difference in terms of anatomical site and source
of infection between TP and SP patients.
Detailed Characteristics of TP patients
Detailed patient characteristics of TP patients are summarized
in Table 2. In only one out of the 15 patients (6.7%), TP was
diagnosed non-operatively. In this patient (patient #14), the
diagnosis of TP was made 5 days after the IO by clinical
signs of infection and laboratory and CT radiographic
J Gastrointest Surg (2009) 13:1358–1367 1361

Table 1 Demographic Data of Patients with Secondary Peritonitis Who Further Developed Tertiary Peritonitis (TP Patients) and Who Did not
Develop Tertiary Peritonitis (SP Patients)

SP patients TP patients

Patients 78.3% (n=54) 21.7% (n=15)

Female 53.7% (n=29) 60.0% (n=9) n.s.
Male 46.3% (n=25) 40.0% (n=6) n.s.
Median age (range), years 72 (14–93) 76 (37–96) n.s.
Mean age (±SD), years 67.1 (±18.3) 70.0 (±18.6) n.s.
Mean BMI (±SD) 25.0 (±5.7) 25.0 (±3.4) n.s.
Cardiovascular co-morbidity (%) 74.1 73.3 n.s.
Pulmonary co-morbidity (%) 38.9 53.3 n.s.
Malignancy (%) 22.2 13.3 n.s.

BMI body mass index, SD standard deviation, n.s. not significant

measurements. In the remaining 14/15 patients (93.3%), TP
was diagnosed intra-operatively by relaparotomies after the
IO (either first or second relaparotomy; Table 2). As required
by the ICU consensus conference definition of TP, these
patients showed persistent or recurrent peritonitis ≥48 h
following successful and adequate surgical source control
which was achieved during IO.2 There was no failure of
surgical source control of the IO (e.g., insufficiency of the
rectal stump, anastomotic insufficiency, etc.). The median
time period between initial operation and diagnosis of TP
was 87 h (range 48–338 h).
Severity, Clinical Course, and Outcome of Secondary
and Tertiary Peritonitis
The mean Mannheim Peritonitis Index, which was recorded at
the IO in all patients (n=69), revealed significant higher values
for TP patients (28.6±SD 7.0; median 20, range 17–39)
compared to SP patients (19.8±SD 8.2; median 20, range 4–
37; p≤0.001, Mann–Whitney test) as illustrated in Fig. 3.
Elevated severity of peritonitis at the IO of TP patients was
paralleled by a higher frequency of relaparotomies following
the IO (14/15 patients; 93.3%) compared to SP patients (5/54
patients; 9.3%; p≤0.001; Fisher’s exact test; Table 3). The
mean number of relaparotomies following IO per patient was
2.00 (±0.93 SD) for TP patients compared to 0.11 (±0.37 SD)
for SP patients (p≤0.001; Mann–Whitney test; Table 3). All
relaparotomies in the five SP patients were “programmed
relaparotomies”. In the TP group, there were nine patients
with “programmed relaparotomies” and five patients with “on
demand relaparotomies” that were initiated by clinical
detection. As a consequence, the concept of “programmed
relaparotomies” was applied with a significantly higher
frequency in TP patients (60.0%) compared to SP (9.3%;
p≤0.001; Fisher’s exact test; Table 3). The timing and
chronology of relaparotomies in relation to the IO is illustrated

Figure 2 Etiology and infection

source of secondary peritonitis
found at the initial operation for
patients who further developed
tertiary peritonitis (TP; n=15)
and for patients who did not
(SP; n=54). Definitions of TP
and SP are according to the ICU
consensus conference.
1362 J Gastrointest Surg (2009) 13:1358–1367

Table 2 Detailed Clinical Data for Patients with Tertiary Peritonitis (n=15)

Patient Age Sex Diagnosis at initial Initial operation (IO) MPI Num. 1st 2nd Diag. ICU f/u (days)
no. (years) operation (IO) at IO relap. relap relap TP (h) stay
(h) (h) (days)

1 85 f Perforated diverticulitis Hartmann’s procedure + 33 2 56 96 56 30 Died (30)

with ileocecal abscess ileocecal resection
2 76 f Rectum perforation and Subtotal colectomy 35 1 48 n.a. 48 3 Died (5)
ischemic ileocecal region with terminal ileostomy
3 80 f Perforation of the ascending Right hemicolectomy with 23 4 41 87 87 77 Surv.
colon terminal ileostomy and
colostomy (mucous fistula)
4 89 f Gastric ulcer perforation Gastric resection (Billroth II) 33 2 41 233 233 10 Died (10)
5 80 f Perforated diverticulitis Hartmann’s procedure + 37 2 338 386 338 17 Died (17)
with multiple interenteric 2 small bowel resections
abscesses and small bowel with primary anastomoses
perforations
6 50 m Perforated diverticulitis Hartmann’s procedure 23 2 42 144 144 16 Surv.
with interenteric abscesses
7 37 m Perforated appendicitis Open appendectomy 20 2 36 90 90 10 Surv.
8 83 f Perforated diverticulitis Hartmann’s procedure 35 2 52 120 52 6 Died (6)
9 72 m Colostomy perforation Segmental resection 23 2 42 89 89 27 Surv.
following parastomal of descending colon,
hernia repair colostomy redo
10 67 f Ileal perforation following Loop ileostomy 39 2 49 99 49 36 Died
subtotal colectomy and (36)
ileo-rectal anastomosis
(anastomosis intact)
h11 48 f Ileal perforation following Closure of perforation, 17 3 58 131 58 10 Surv.
anterior rectum resection lavage
(anastomosis intact)
12 37 m Small bowel perforation Ileocecal resection, 20 3 36 96 96 13 Surv.
due to briden ileus, loop ileostomy
Crohn’s disease
13 70 f Perforated diverticulitis Hartmann’s procedure 29 2 86 264 86 34 Died (38)
14 80 m Gastric perforation Closure of perforation 32 0 n.a. n.a. 120a 9 Died (9)
due to advanced
gastric cancer
15 96 m Perforated Hartmann’s procedure 30 1 60 n.a. 60 2 Died (3)
diverticulitis

f female, m male, IO initial operation, MPI Mannheim Peritonitis Index, Num. Relap. number of relaparotomies, 1st/2nd relap Time period
between initial operation and first/second relaparotomy in hours, Diag. TP time period between initial operation and diagnose of tertiary
peritonitis (TP), f/u follow-up, Surv. patient still alive, Died (x) patient died x days after the initial operation in the hospital
a
Diagnosis of TP was made based in clinical and laboratory findings only (patient # 14)

in Fig. 4. Impaired outcome of TP patients compared to SP Microbiological Data

patients was paralleled by significantly longer hospitalization
on the intensive care unit, since median intensive care unit
stay for TP patients was 13 days (range 3–77 days) compared
to 4 days (range 1–50 days) for SP patients (p=0.002, Mann–
Whitney test; Table 3). Compared to SP patients, TP patients
were characterized by higher frequency of multi-organ failure
(73.3% vs. 18.5%; p≤0.001, Fisher’s exact test) and higher
mortality (60.0% vs. 9.3%; p≤0.001, Fisher’s exact test;
Table 3). All deaths in the TP group (9/15) were due to septic
multi-organ failure as a result of tertiary peritonitis. There
were no autopsies performed.
Figures 5 and 6 illustrate the microbiological spectrum of
microbial isolates obtained from the IO and the antibiotic
therapy initiated during the IO—separately for TP patients
(n=15 specimens) compared to SP patients (n=54 speci-
mens). The distribution of microbiological species at the time
of the IO did not differ significantly between TP and SP
patients with the exception of Escherichia coli. There was a
significantly higher proportion of E. coli in isolates from TP
patients compared to SP patients (73.3% vs. 37.0%; p≤0.05;
Fisher’s exact test; Fig. 5). As depicted in Fig. 6, antibiotic
J Gastrointest Surg (2009) 13:1358–1367 1363

(265 ± 17 vs. 217± 12 mg/L; p =0.05, T test) and on

Figure 3 Mannheim Peritonitis Index (MPI) at the initial operation.
Significantly higher MPI values for patients who further developed
tertiary peritonitis (TP; n=15) compared to patients who did not (SP;
n=54; p≤0.001; Mann–Whitney test). Boxes represent the lower,
median, and upper quartiles; whiskers indicate the 10th–90th
percentile and outliers are illustrated by dots.
therapy initiated during IO after detection of secondary
peritonitis did not differ significantly between TP and SP
patients. The majority of patients were treated with imipenem/
cilastatin (SP 27.8%; TP 46.7%; p=0.21; Fisher’s exact test)
or piperacillin/tazobactam (SP 51.9%; TP 53.3%; p=1.00;
Fisher’s exact test). Figure 7 delineates the changes in the
microbiological spectrum in TP patients compared to SP
patients. This analysis compares isolates of TP patients from
the relaparotomy that was diagnostic for TP and isolates from
the IO of SP patients (n=54 specimens). In the TP group
(n=15 patients), only 11 specimens were obtained during
relaparotomy and could be included into the analysis. There
was a significant microbiological shift towards Enterococcus
and Candida species in TP with significantly higher propor-
tions of Enterococcus (*p≤0.05; Fisher’s exact test) and
Candida (**p≤0.01; Fisher’s exact test) in TP patients
compared to SP patients (Fig. 7).
postoperative day 7 (174±23 vs. 119±11 mg/L; p=0.03,
T test; Fig. 8).
The mean SAPS II score (±SEM) during the first three
postoperative days after the IO operation was significantly
higher in TP patients (46.1±3.7) compared to SP patients
(29.7±2.0) (p≤0.001, T test). The time course of SAPS II
values during the first three postoperative days after the
initial operation is depicted in Fig. 9. SAPS II scores for TP
patients on the first (47.1±4.2), second (45.1±4.0), and
third postoperative days (44.9±4.0) were significantly
higher compared to SP patients on the respective days
(30.7±2.1, 28.4±2.0, and 30.3±2.5, respectively; p≤0.001,
p≤0.001, and p=0.004, respectively; T test; Fig. 9).
Early Detection of Tertiary Peritonitis
In order to asses to what extent intra-operative Mannheim
Peritonitis Index measurement during the IO and C-reactive
protein and SAPS II measurements on postoperative day 2
could differentiate between TP patients and SP patients, the
corresponding receiver operating characteristic curve was
constructed and the area under the receiver operator
characteristic curve was calculated. The area under the
receiver operator characteristic curve for the Mannheim
Peritonitis Index at the initial operation was 0.794 (95%
confidence interval=0.672–0.915; p≤0.001). A sensitivity
of 80.0% and specificity of 68.5% were achieved with a
Mannheim Peritonitis Index cut-off value of 22 (Table 3).
The area under the receiver operator characteristic curve for
C-reactive protein and SAPS II on the second postoperative
Table 3 Clinical Course and Outcome of Patients with Secondary and
Tertiary Peritonitis
Secondary Tertiary Peritonitis
Peritonitis

Patients 78.3% (n=54) 21.7% (n=15)

Laboratory Parameters
Frequency of 9.3% (n=5) 93.3% (n=14) p≤0.001
relaparotomy
The mean C-reactive protein (±SEM) during the first Relaparotomy/ 0.11 (±0.37) 2.00 (±0.93) p≤0.001
postoperative days after the IO (postoperative day 1– patient (±SD)
Frequency of 9.3% (n=5) 60.0% (n=9) p≤0.001
postoperative day 7) was significantly higher in TP
“programmed”
patients (204±13 mg/L) compared to SP patients (166± relaparotomy
8 mg/L; p≤0.05, T test). The time course of C-reactive Frequency of 0% (n=0) 33.3% (n=5) p≤0.001
protein values during the first postoperative days after the “on demand”
relaparotomy
IO is displayed in Fig. 8 for SP and TP patients. Both
Median ICU 4 days (1–50) 13 days p=0.002
curves decline from preoperative values to postoperative stay (range) (3–77 years)
day 1. On the second postoperative day, C-reactive protein Frequency of 18.5% 73.3% p≤0.001
is at its maximum and again declining over the next days. MOF
Mortality 9.3% 60.0% p≤0.001
Although both curves run parallel to each other, mean C-
reactive protein values for TP patients are significantly higher SD standard deviation, ICU intensive care unit, MOF multiple organ
compared to SP patients on the second postoperative day failure
1364 J Gastrointest Surg (2009) 13:1358–1367

Figure 4 Timing and chronolo-

gy of the first relaparotomy
(1st Relap) and second
relaparotomy (2nd Relap) for
patients who further developed
tertiary peritonitis (TP; n=15)
compared to patients who did
not (SP; n=54) in relation to the
initial operation. Each dot
represents an individual patient.

day after initial operation was 0.696 (95% confidence
interval=0.562–0.830; p=0.02) and 0.797 (95% confidence
interval=0.634–0.960; p≤0.001), respectively. A cut-off
value for C-reactive protein of 215 mg/L led to a sensitivity
of 80.0% and a specificity of 57.4%. A cut-off value of 39
for the SAPS II score revealed a sensitivity of 80.0% with a
specificity of 74.5% (Table 4).
Discussion
Definition of Tertiary Peritonitis
The standard treatment for SP is an immediate laparotomy
with surgical source control and antibiotic therapy. How-
ever, a few patients will develop a clinical syndrome—also
referred to as TP, which is characterized by a persistent
intra-abdominal infection, an altered microbial flora, failure
of the immune response, and progressive organ dysfunction
leading to high mortality. There is still an ongoing debate
about the definition of TP. In fact, some opinions deny the
existence of TP as a distinct entity. In the past, TP has
simply been defined as failed surgical source control or
inadequate antibiotic therapy of SP. Other definitions
emphasized the impaired host response to peritoneal
infection.16 This heterogeneity of definitions resulted in
varying inclusion criteria and incommensurable results in
clinical studies focusing on TP.5 In the current study, we
applied the latest ICU consensus conference guideline that
provides a precise definition. TP was defined as intra-
abdominal infection that persists or recurs ≥48 h following
successful and adequate surgical source control.2 This
definition contains two essential conditions, which have to
be met: the time period (≥48 h) and successful surgical
source control. Although the ICU guideline does not
provide further explanation for “successful surgical source
control”,2 our interpretation of this term was a complete and
sustainable eradication of the surgical focus. If a patient

Figure 5 Microbiological
spectrum of microbial isolates
obtained from the initial
operation. Comparison between
patients who further developed
tertiary peritonitis (TP; n=15
specimens) compared to patients
who did not (SP; n=54
specimens). Significantly higher
proportion of E. coli in TP
compared to SP (*p≤0.05;
Fisher’s exact test). Dotted lines
separate gram-negative bacteria,
gram-positive bacteria, and
fungi.
J Gastrointest Surg (2009) 13:1358–1367
Figure 6 Antibiotic treatment initiated during the initial operation at
the onset of secondary peritonitis. There was no difference in the
antibiotic spectrum between patients who further developed tertiary
peritonitis (TP; n=15) compared to patients who did not (SP; n=54
specimens; Fisher’s exact test).
presented—during relaparotomy or clinically—with an
obvious failure of previous surgical source control in terms
of a “technical problem”, this patient was not classified as
TP but as SP patient. Other examples of “failure of surgical
source” control comprise insufficiency of the rectal stump
after Hartmann’s procedure, anastomotic insufficiency, or
other technical problems that lead to disruption of the
physical integrity of the gastrointestinal hollow organs.
Nevertheless, there is consensus that SP and TP exist in
a continuum and the transition between both may be quite
subtle. Although TP may be diagnosed during relaparotomy
as a simple discrete point in the illness, in reality, it evolves
gradually over several hours or days. In the current study,
TP was diagnosed during relaparotomy in 14/15 patients.
Only one patient was diagnosed having TP by clinical and
laboratory measures 120 h after initial operation. For all
Figure 7 Comparison of the
microbiological spectrum
between secondary peritonitis
(SP) and tertiary peritonitis (TP).
The microbial isolates of patients
with TP were obtained from the
relaparotomy that was diagnostic
for TP (n=11 specimens).
Isolates of patients with SP were
obtained from the initial
operation (n=54 specimens).
Significantly higher proportion of
Enterococcus (*p≤0.05; Fisher’s
exact test) and Candida (**p≤
0.01; Fisher’s exact test) in TP
compared to SP. Dotted lines
separate gram-negative bacteria,
gram-positive bacteria, and fungi.
1365
patients with TP, the time interval between the initial
operation and the diagnosis of TP was 87 h (median) and
thus considerably long. In addition, it is important to
emphasize that in six patients the diagnosis was made not
until the second relaparotomy, while during the first
relaparotomy the intra-abdominal situation was estimated
innocuously. It was therefore the aim of this study to
compare clinical and laboratory parameters between
patients with SP who will further develop TP (TP patients)
and who will not (SP patients). The necessity to define early
predictors for TP becomes evident looking upon the devas-
tating mortality rate for TP of 60% encountered in this study,
which was relatively high compared to other studies—
reporting mortality rates ranging between 27% and
64%.11,12,17 We also observed a clear relationship between
peritonitis type (TP vs. SP) and mortality, which was in
contrast to other publications.16
Risk Factors and Microbial Flora of TP
Several epidemiologic and clinical risk factors have already
been identified that might predispose to TP, which include
age, etiology of peritonitis, malnutrition, and multi-resistant
microorganisms.15 With regard to the patient’s age or
etiology and infection source of peritonitis, we were unable
to detect significant differences between TP and SP.
Concerning the microbial flora encountered in the initial
operation, we did only find a higher proportion of E. coli in
TP patients compared to SP patients. All other bacteria
were equally distributed. It has recently been shown that
there is a microbial shift in TP towards Enterococcus,
Enterobacter, Pseudomonas, Candida albicans and other
opportunistic bacteria and fungi.11,12,17 However, in this
study, we could only demonstrate a significant shift towards
Enterococcus and C. albicans between patients who
suffered from TP compared to SP. In our opinion,
1366
Figure 8 Time course of C-reactive protein (CRP) in the perioper-
ative period of the initial operation in patients who further developed
tertiary peritonitis (TP) and patients who did not (SP). Mean CRP ±
SEM values are indicated preoperatively (preop.) and on postoperative
days 1, 2, 3, and 7 (d1, d2, d3, and d7). Significantly higher CRP
values for TP compared to SP on the second postoperative day
(p=0.05) and postoperative day 7 (p=0.03; T test).
microbiology is not suited as an early diagnostic marker for
the identification of patients at risk for TP, since microbi-
ological studies—including resistance analysis—take up to
1 week. Nevertheless, future studies will be necessary to
investigate the microbial shift as well as the antibiotic
resistance data in our patients.
Predictive Parameters for TP
In the current study, we analyzed three early and easily
accessible parameters for identification of patients who might
further develop TP: Mannheim Peritonitis Index, SAPS II, and
C-reactive protein. Some might argue that due to persisting
systemic inflammation repeated surgical procedures or inter-
mittent nosocomial infections, the value of clinical (Mannheim
Figure 9 Time course of SAPS II scores in the postoperative period
after the initial operation in patients who further developed tertiary
peritonitis (TP) and patients who did not (SP). Mean SAPS II scores ±
SEM values are indicated on the first three postoperative days
(d1–d3). Significantly higher SAPS II scores for TP compared to SP
during the whole period (p≤0.001, p≤0.001, and p=0.004, respec-
tively; T test).
J Gastrointest Surg (2009) 13:1358–1367
Table 4 Diagnostic Accuracy of MPI at Initial Operation and CRP/
SAPS II on the Second Postoperative Day for the Discrimination
Between Tertiary Peritonitis and Secondary Peritonitis
Cut-off value Sensitivity (%) Specificity (%) LR+
MPI 22 80.0 [51.9–95.7] 68.5 [54.5–80.5] 2.54
CRP 215 (mg/L) 80.0 [51.9–95.7] 57.4 [43.2–70.8] 1.88
SAPS II 39 80.0 [51.9–95.7] 74.5 [59.7–86.1] 3.13
Values in square brackets are 95% confidence interval
MPI Mannheim Peritonitis Index, CRP C-reactive protein (milligram
per liter), MPI Mannheim Peritonitis Index, SAPS II Simplified Acute
Physiology Score II, LR+ positive likelihood ratio
Peritonitis Index, SPAS II) and laboratory parameters (C-
reactive protein) for sufficient diagnosis of TP is limited.5 In
fact, there are conflicting data concerning the value applying
such parameters for the detection of TP.15,17 Unlike other
studies, our approach was to analyze these parameters as early
as possible—at the IO that was diagnostic for SP and on the
first postoperative days.
The Mannheim Peritonitis Index was initially designed to
estimate the prognosis and predict mortality of patients with
SP.8–10 In our study population, the Mannheim Peritonitis
Index was significantly higher in patients that later on
developed TP compared to SP (28.6 vs. 19.8). Similar results
have been shown in two recent publications analyzing the
Mannheim Peritonitis Index in TP.11,12 In addition, the
receiver operator characteristic analysis in the current study
revealed encouraging results with an area under the receiver
operator characteristic curve of 0.794 for the detection of TP.
With regard to the receiver operator characteristic analysis, it
has to be considered that the Mannheim Peritonitis Index is
an early—if not the earliest—marker for TP. It is accessible
immediately during the IO. This renders the Mannheim
Peritonitis Index to a diagnostic tool of high potential.
The second parameter was the SAPS II score, initially
designed to predict mortality and disease severity of critical
ill patients.13,14 We could demonstrate that SAPS II was
significantly higher during the first three postoperative days
after initial operation in TP patients (46.0) compared to SP
patients (29.7). Interestingly, the curves for TP and SP
patients ran completely parallel to each other over the
whole period. The receiver operator characteristic analysis
on the second day revealed an area under the receiver
operator characteristic curve of 0.797, which demonstrates
the diagnostic potential of this scoring system for early
identification of patients at risk for TP. Our results are
consistent with a recent study that reported similar SAPS II
scores for TP (45.6) and SP (31.9) patients—underlining
the importance of this parameter.12
The third parameter tested in our study was the acute
phase protein C-reactive protein. Although C-reactive
J Gastrointest Surg (2009) 13:1358–1367
protein constitutes a routine parameter in patients with
abdominal infections, it has hardly been explicitly evaluated
in the diagnosis of TP.5,15 In our study, the time course of C-
reactive protein displayed a curve with two peaks: one peak
preoperatively and one peak on the second postoperative day
after the IO. In between, on the first postoperative day, lower
C-reactive protein values were observed, possibly due to an
operative clearing effect. Interestingly, although both curves
run parallel, C-reactive protein values of TP patients were
significantly higher compared to SP patients on the peak of
the second postoperative day (265 vs. 217) after the IO.
However, the corresponding area under the receiver operator
characteristic curve was only 0.696. The main problem of C-
reactive protein is the lack of specificity for abdominal
infections, as shown in numerous studies.18–20 A rise of C-
reactive protein during the postoperative period may simply
be the result of the operative trauma.21,22 Nevertheless, this
study shows a high diagnostic potential of C-reactive protein.
This hypothesis has to be addressed in further studies.
Conclusion
In conclusion, due to high mortality of tertiary peritonitis and
often delayed diagnosis, it is crucial to identify patients at risk
for developing tertiary peritonitis as early as possible: at the
initial operation that reveals the diagnosis of peritonitis and
during the first postoperative days. Our results indicate that
the Mannheim Peritonitis Index assessed at the initial
operation and the time course of C-reactive protein and SAPS
II during the first days after initial operation are promising
diagnostic candidates for the future.
Acknowledgments The authors thank all the members of the
intensive care unit team for continuous support. They further thank
Tim Holland-Letz, Department of Medical Informatics, Biometry and
Epidemiology, Ruhr-University Bochum for statistical assistance.
References
1. Weiss G, Steffanie W, Lippert H. Peritonitis: main reason of
severe sepsis in surgical intensive care. Zentralbl Chir
2007;132:130–137. doi:10.1055/s-2006-960478.
2. Calandra T, Cohen J. The international sepsis forum consensus
conference on definitions of infection in the intensive
care unit. Crit Care Med 2005;33:1538–1548. doi:10.1097/
01.CCM.0000168253.91200.83.
3. Reemst P, Van Goor H, Goris RJ. SIRS, MODS and tertiary
peritonitis. Eur J Surg 1996;576(Suppl):47–48. discussion 49.
4. Buijk SE, Bruining HA. Future directions in the management of
tertiary peritonitis. Intensive Care Med 2002;28:1024–1029.
doi:10.1007/s00134-002-1383-6.
5. Evans HL, Raymond DP, Pelletier SJ, Crabtree TD, Pruett TL,
Sawyer RG. Diagnosis of intra-abdominal infection in the
1367
critically ill patient. Curr Opin Crit Care 2001;7:117–121.
doi:10.1097/00075198-200104000-00010.
6. Koperna T, Schulz F. Relaparotomy in peritonitis: prognosis and
treatment of patients with persisting intraabdominal infection.
World J Surg 2000;24:32–37. doi:10.1007/s002689910007.
7. Koperna T. Surgical management of severe secondary peritonitis. Br
J Surg 2000;87:378. doi:10.1046/j.1365-2168.2000.01369-12.x.
8. Linder MM, Wacha H, Feldmann U, Wesch G, Streifensand RA,
Gundlach E. The Mannheim Peritonitis Index. An instrument for
the intraoperative prognosis of peritonitis. Chirurg 1987;58:84–
92.
9. Billing A, Frohlich D, Schildberg FW. Prediction of outcome
using the Mannheim Peritonitis Index in 2003 patients. Peritonitis
Study Group. Br J Surg 1994;81:209–213. doi:10.1002/
bjs.1800810217.
10. Fugger R, Rogy M, Herbst F, Schemper M, Schulz F. Validation
study of the Mannheim Peritonitis Index. Chirurg 1988;59:598–
601.
11. Panhofer P, Izay B, Riedl M, Ferenc V, Ploder M, Jakesz R,
Gotzinger P. Age, microbiology and prognostic scores help to
differentiate between secondary and tertiary peritonitis. Langen-
becks Arch Surg 2009;394(2):265–271.
12. Weiss G, Meyer F, Lippert H. Infectiological diagnostic problems
in tertiary peritonitis. Langenbecks Arch Surg 2006;391:473–482.
doi:10.1007/s00423-006-0071-3.
13. Agha A, Bein T, Frohlich D, Hofler S, Krenz D, Jauch KW.
“Simplified Acute Physiology Score” (SAPS II) in the assessment
of severity of illness in surgical intensive care patients. Chirurg
2002;73:439–442. doi:10.1007/s00104-001-0374-4.
14. Le Gall Jr., Lemeshow S, Saulnier F. A new Simplified Acute
Physiology Score (SAPS II) based on a European/North American
multicenter study. JAMA 1993;270:2957–2963. doi:10.1001/
jama.270.24.2957.
15. Malangoni MA. Evaluation and management of tertiary peritoni-
tis. Am Surg 2000;66:157–161.
16. Evans HL, Raymond DP, Pelletier SJ, Crabtree TD, Pruett TL,
Sawyer RG. Tertiary peritonitis (recurrent diffuse or localized
disease) is not an independent predictor of mortality in surgical
patients with intraabdominal infection. Surg Infect (Larchmt)
2 0 0 1 ; 2 : 2 5 5 – 2 6 3 . d is c u s s i o n 2 6 4 – 2 5 5 . d o i : 1 0 . 1 0 8 9 /
10962960152813296.
17. Nathens AB, Rotstein OD, Marshall JC. Tertiary peritonitis:
clinical features of a complex nosocomial infection. World J Surg
1998;22:158–163. doi:10.1007/s002689900364.
18. Dahaba AA, Hagara B, Fall A, Rehak PH, List WF, Metzler H.
Procalcitonin for early prediction of survival outcome in postop-
erative critically ill patients with severe sepsis. Br J Anaesth
2006;97:503–508. doi:10.1093/bja/ael181.
19. Mokart D, Merlin M, Sannini A, Brun JP, Delpero JR,
Houvenaeghel G, Moutardier V, Blache JL. Procalcitonin,
interleukin 6 and systemic inflammatory response syndrome
(SIRS): early markers of postoperative sepsis after major surgery.
Br J Anaesth 2005;94:767–773. doi:10.1093/bja/aei143.
20. Reith HB, Mittelkotter U, Debus ES, Kussner C, Thiede A.
Procalcitonin in early detection of postoperative complications.
Dig Surg 1998;15:260–265. doi:10.1159/000018625.
21. Lindberg M, Hole A, Johnsen H, Asberg A, Rydning A, Myrvold
HE, Bjerve KS. Reference intervals for procalcitonin and C-
reactive protein after major abdominal surgery. Scand J Clin Lab
Invest 2002;62:189–194. doi:10.1080/003655102317475443.
22. Meisner M, Tschaikowsky K, Hutzler A, Schick C, Schuttler J.
Postoperative plasma concentrations of procalcitonin after differ-
ent types of surgery. Intensive Care Med 1998;24:680–684.
doi:10.1007/s001340050644.